This subproject is one of many research subprojects utilizing the resources provided by a Center grant funded by NIH/NCRR. The subproject and investigator (PI) may have received primary funding from another NIH source, and thus could be represented in other CRISP entries. The institution listed is for the Center, which is not necessarily the institution for the investigator. BACKGROUND: Although the genomes of many of the most important human and animal pathogens have now been sequenced, our understanding of the actual proteins expressed by these genomes and how well they predict protein sequence and expression is still deficient. We have used three complementary approaches (two-dimensional electrophoresis, gel-liquid chromatography linked tandem mass spectrometry and MudPIT) to analyze the proteome of Toxoplasma gondii, a parasite of medical and veterinary significance, and have developed a public repository for these data within ToxoDB, making for the first time proteomics data an integral part of this key genome resource. RESULTS: The draft genome for Toxoplasma predicts around 8,000 genes with varying degrees of confidence. Our data demonstrate how proteomics can inform these predictions and help discover new genes. We have identified nearly one-third (2,252) of all the predicted proteins, with 2,477 intron-spanning peptides providing supporting evidence for correct splice site annotation. Functional predictions for each protein and key pathways were determined from the proteome. Importantly, we show evidence for many proteins that match alternative gene models, or previously unpredicted genes. For example, approximately 15% of peptides matched more convincingly to alternative gene models. We also compared our data with existing transcriptional data in which we highlight apparent discrepancies between gene transcription and protein expression. CONCLUSION: Our data demonstrate the importance of protein data in expression profiling experiments and highlight the necessity of integrating proteomic with genomic data so that iterative refinements of both annotation and expression models are possible.